RedHatAI / text-generation updated 1 year ago

Llama-3.2-3B-Instruct-FP8-dynamic

- Model Architecture: Meta-Llama-3.2 - Input: Text - Output: Text - Model Optimizations: - Weight quantization: FP8 - Activation quantization: FP8 - Intended Use Cases: Intended for commercial and research use in multiple languages. Similarly to Llama-3.2-3B-Instruct, this models is intended for assistant-like chat. - Out-of-scope: Use in any manner that violates applicable laws or regulations (including trade compli...

Params
3.6 B
Context
131,072
Downloads 30d
277 K
Likes
3
Commercial use: conditional · llama3.2 Not gated SAFETENSORS 8 languages View on Hugging Face ↗

Download history

tracking started — chart appears after 7 days of snapshots (5 recorded)
277 K downloads in the last 30 days

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

FileQuantSizeEst. VRAMVerdict on RTX 4090 · 24 GB
model.safetensors f8_e4m3 4.4 GB 5.9 GB ✅ Runs comfortably
model.safetensors (bf16, full) bf16 + 128K ctx 4.4 GB 14.0 GB ✅ Runs comfortably
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Run it

copy-paste, exact tags checked against the Hub
~ · curl · api/v1
$ curl -s https://aimodelscomparison.com/api/v1/models/llama-3-2-3b-instruct-fp8-dynamic
{
  "hf_id": "RedHatAI/Llama-3.2-3B-Instruct-FP8-dynamic",
  "params_b": 3.61,
  "context_length": 131072,
  "license": { "id": "llama3.2", "commercial": "conditional" },
  "downloads_30d": 276876,
  "vram_estimates": [
    { "quant": "f8_e4m3", "gb": 5.9 }
  ],
  "updated_at": "2026-09-17T01:00:34Z"
}
est. VRAM —on RTX 4090 · 24 GBJSON API →

Specifications

Architecture
LlamaForCausalLM
Parameters
3.6 B
Tensor type
F8_E4M3
Context length
131,072
Vocabulary
128,256
Layers / heads
28 / 24
Licence
llama3.2
First seen on the Hub
2024-09-25
Base model
Llama-3.2-3B-Instruct
Training datasets
undisclosed
Added to our catalog
2026-09-17

Family

Base model and the most-downloaded derivatives in the catalog.

Compare with any text-generation model